User guide

sports-betting extracts sports betting data and trains predictive models on it. You need to know two objects. A dataloader gets the data. A bettor bets on it.

Dataloader

Betting data rarely comes in a shape you can model directly. The dataloader extracts it in a consistent format. Here is a dataloader for the Italian and Spanish leagues, seasons 2023 and 2024.

from sportsbet.dataloaders import DataLoader
from sportsbet.sources import FootballDataOdds, FootballDataStats
dataloader = DataLoader(
    param_grid={'league': ['Italy', 'Spain'], 'year': [2023, 2024]},
    stats=FootballDataStats(),
    odds=FootballDataOdds(),
)

Extract the training data with the market maximum odds.

X_train, Y_train, O_train = dataloader.extract_train_data(odds_type='market_maximum')

Then extract the fixtures data.

X_fix, Y_fix, O_fix = dataloader.extract_fixtures_data()

Bettor

You now have the training and fixtures data. A bettor evaluates a model and predicts the value bets of the upcoming matches. Here is a ClassifierBettor around a scikit-learn KNeighborsClassifier.

from sportsbet.evaluation import ClassifierBettor, backtest
from sklearn.neighbors import KNeighborsClassifier
from sklearn.impute import SimpleImputer
from sklearn.pipeline import make_pipeline
bettor = ClassifierBettor(classifier=make_pipeline(SimpleImputer(), KNeighborsClassifier()))

Backtest it on the historical data, using the numerical features.

num_cols = X_train.columns[['float' in col_type.name for col_type in X_train.dtypes]]
backtest(bettor, X_train[num_cols], Y_train, O_train)

Fit it and predict the value bets of the fixtures.

bettor.fit(X_train[num_cols], Y_train)
value_bets = bettor.bet(X_fix[num_cols], O_fix)